|
|
Gelman, Andrew, Jennifer Hill, and Aki Vehtari. 2020. Regression and other stories. Cambridge: Cambridge University Press. ROS |
Free to download PDF version from the book’s website: https://avehtari.github.io/ROS-Examples |
|
|
Alexander, Rohan. 2023. Telling Stories with Data. Chapman and Hall/CRC TSD |
Free online book: https://tellingstorieswithdata.com |
|
|
Gelman, A., and Hill, J. 2007. Data Analysis Using Regression and Multilevel/Hierarchical Models. Cambridge: Cambridge University Press. ARM |
Note: ROS is the expanded and updated version of Part 1 (and some of Part 3) of this book. While everyone in the free world eagerly awaits the publication of ROS’s multilevel counterpart, we’ll use ARM as a reference work for the theory underpinning multilevel modelling. Not freely available. Access it in print or online via the NU library |
Textbooks
The course does not strictly follow the content of a textbook, but the expectation is that students will read as much as possible of the assigned chapters from the following books:
Relatively large portions of text will be assigned for reading in each week from these books, referring to them by their acronyms. Don’t worry if you cannot read all the textbook content assigned in any given week! Those for whom the method covered by the assigned readings is new, will be able to refer back to them throughout the semester and beyond, reading thoroughly and completing the applied exercises. Those already familiar to some extent with the methods, should nonetheless read the text as a narrative and will discover hidden gems that will spectacularly improve their understanding and ability to interpret their statistical results.
Application
In the IT labs we will practice applying methods by reproducing small bits of published research, using the data and (critically) the modelling approaches used by the authors. To fully understand the context of these data and the methods used, you must read the original journal articles and the available supplementary materials provided alongside. These readings will be listed under each week’s outline (still work in progress!).
The articles come from a variety of different fields, so expect them to push you outside your disciplinary comfort zone. The point is to see how methods have been used in practice and learn how to reproduce (and potentially improve) those analyses. This will then enable you to apply this knowledge to your own research questions.
When selecting the articles, the aim was to strike a fine balance between (a) the simplicity of the methods employed, (b) data and analytical transparency, and (c) the strength of the analysis. So don’t take them as examples of all-rounds best practice, but examples of research that gets published while being self-confident enough to open itself up for public scrutiny. Aim for this in your own research!
Technique
There will also be various readings relating more closely to the technicalities of coding in R and scientific writing, collaboration and communication in general. These readings will also be listed under each week’s outline as the semester progresses. The generic reading that students are advised to go through on their own is:
|
|
Wickham, Çetinkaya-Rundel and Grolemund. 2022. R for Data Science (2nd ed.) R4DS |
Free online book: https://r4ds.hadley.nz/ |
Intuition
Finally, there will also be recommended readings listed under certain weeks that help place methods, statistics and probability theory in a broader frame. These are useful readings for everyone, regardless of whether you will be applying quantitative analysis in your research or future work.
Required software
We will use a number of open-source software for data analysis and scientific writing. You need to install these on your personal computers to be able to work away from campus:
Students with no previous experience using R and/or RStudio are advised to complete the self-paced free online training course R for Social Scientists provided by Data Carpentry at https://datacarpentry.org/r-socialsci/
There are several ways to get help with outside class. If you encounter an error message or are looking for a function to perform a specific task that we have not covered in class, you can do a Google search; for best results, use the https://rseek.org/ search engine, which limits the results to those relating to the language.
You can also search for answers on Stack Overflow, which is a popular help and discussion website for programmers. You can also post a question there, but make sure to follow community standards and advice on how to ask a good question and how to provide a minimal reproducible example. You will need some experience using the site before being able to ask a good question, but it’s more than certain that any question you have at this stage has already been asked and answered somewhere. Make sure you do a comprehensive search with various prompts before thinking about asking your question.
Increasingly, large language model-based chatbots such as the (in)famous ChatGPT can also provide good answers. You can use them efficiently, but make sure to always test out the responses, in the overwhelming majority of the cases the advice they give is unreliable, at least at first try.